The AI Reckoning Has Turned Violent, and Nobody Wants to Talk About Why

---

A man drove to Sam Altman’s home with documents advocating violence against AI executives. He’s now facing federal felony charges and attempted murder.

That’s not a thought experiment. That’s not a Reddit thread. That’s 2025.

We’re watching the collision between three colliding realities all at once: AI is replacing jobs faster than anyone predicted, tech companies are making decisions that feel fundamentally inhuman, and the social fabric that used to absorb economic shock is fraying. The violence against Altman is the canary in the coal mine—and we’re still pretending it’s singing normally.

Photo by Jorge Soto Farias / Pexels

The Productivity Gamble

Let’s start with what companies actually want from AI right now. Booking.com just got hacked, and customers had their reservations hijacked. Google’s about to punish websites that trap you with broken back buttons. These aren’t accidents. They’re symptoms of an industry moving at escape velocity, cutting corners because the upside is so massive that the cleanup costs barely register.

Now look at the digital twin trend. Firms are building AI versions of their employees—digital doubles that work 24/7, don’t need breaks, don’t unionize. The pitch is seductive: you become a “superworker.” The reality is messier. Who owns the data? If the digital twin makes a mistake, who’s liable? What happens to your leverage in salary negotiations when there’s a perfect digital copy of you ready to take your job for server costs?

I think this is what venture capital looks like when it’s running out of things to disrupt: turning humans into optimization problems.

Photo by nappy / Pexels

The Job Disappearance, Accelerating

Snap just laid off 16% of its workforce—about 1,000 people—and explicitly said it’s betting on AI to do their work instead. Allbirds, the sneaker company that couldn’t crack retail, sold itself for $39 million and immediately rebranded as “NewBird AI” to buy computer chips. These aren’t isolated decisions. They’re a coordinated signal: the era of “AI will create new jobs” is over. We’re in the era of “AI will eliminate jobs, and we’re okay with that.”

Here’s what fascinates me: the one type of work that’s actually increasing right now? The human stuff. Meetings. Negotiation. Cajoling. Reassurance. The work of keeping humans from fragmenting when everything around them is becoming algorithmic. That’s not a job market—that’s a coping mechanism dressed up as opportunity.

But not everyone copes well. Some people radicalize.

Why Violence Matters

The attack on Altman’s home isn’t an isolated incident. It’s the symptom of something the tech industry has been trained to ignore: cumulative social harm has a breaking point. You can’t tell millions of people their skills are obsolete, their communities are disrupted, their futures are uncertain, and then act shocked when one of them decides violence is the only language that gets heard.

This isn’t a defense of the violence. It’s an explanation. And explanations matter because they’re predictive.

The UK government is now seriously considering banning under-16s from social media. Prime Minister Starmer told tech bosses: “Things can’t go on like this with online safety.” That’s the sound of political institutions finally admitting they’ve lost control. Social media has been fragmenting attention, attention disorder is up, mental health is down, and nobody’s actually fixed it. AI is about to make it worse—more personalization, more addiction, more radicalization, all optimized at scale.

This is the recipe for more violence. Not because AI is evil. Because institutions aren’t equipped to manage the speed of change, and people respond to powerlessness with aggression.

The Actual Insight About AI You Need

There’s a concept floating around called “jagged intelligence”—the idea that AI isn’t a smarter human, it’s something fundamentally different. It’s better at pattern-matching and prediction. Worse at judgment and context. Brilliant at summarizing text. Useless at knowing what a community actually needs.

Here’s what that means: the jobs that disappear first are the ones that look like pattern-matching from the outside but actually require judgment. Customer service, content moderation, data analysis, basic coding. The jobs that survive longest are the ones where you need to understand messy, human, context-dependent situations.

Except here’s the problem: we’re automating the wrong end of the skill spectrum. We’re replacing low-wage work with AI, which sounds humane until you realize there’s no retraining pipeline, no safety net, and the wealth generated goes to capital, not displaced workers. Meanwhile, we’re keeping the jobs that require human judgment—the ones that actually need cognitive work—at the same pay and workload as before.

My read is that the next five years will be defined by whether we build any social infrastructure to absorb this shock. If we don’t, the violence won’t be an outlier. It’ll be a trend.

Photo by Markus Spiske / Pexels

What the Tech Industry Got Wrong

Silicon Valley has always operated on a simple philosophy: move fast and break things. It worked great when the things you were breaking were Blockbuster’s business model or taxi dispatch. It works less well when the things you’re breaking are employment, mental health, and social trust.

The fact that Snap and Allbirds are moving to AI isn’t surprising. The fact that they’re doing it publicly, without apology, with the vague assurance that “new jobs will emerge”—that’s the real story. They’re signaling that they don’t expect pushback. They’re right, so far.

But the attack on Altman changed the calculation slightly. It’s not because one violent incident means anything on its own. It’s because it made the unspoken explicit: some people are radicalized enough to act.

And institutions hate that. It forces a response.

What I’m Watching

The UK under-16 social media ban implementation timeline. This will determine whether democracies can actually regulate tech at scale. If it works, expect copycat legislation in the US and EU within 18 months. If it fails—if it’s unenforceable or gets challenged in court—we learn that regulation is theater.

Booking.com’s disclosure of how many customers were affected by the hack. The platform said it changed security protocols but wouldn’t say the scope. Watch whether regulators force transparency. That number will tell you whether we’re in the age of security theater or actual accountability.

Snap’s Q2 earnings call commentary on whether the layoffs actually improved profitability. If they did, every other social media company will follow. If they didn’t—if they just shifted costs and didn’t actually improve the business—that’s a signal that the AI pivot is hype covering for deeper problems. That changes the narrative.

Whether another AI executive gets attacked in the next 18 months. This is the uncomfortable metric. One attack is an outlier. Two is a pattern. Three is a movement. I hope I’m wrong about this one.